JP4839133B2 - Data management method and computer system for storage apparatus - Google Patents

Description

  The present invention relates to a data management method for a storage apparatus, and more particularly to a data management method when using a differential snapshot technique.

  In a computer system, data stored in a storage device may be duplicated for data backup or data check. As a method of copying data, there is a method of copying all data to another data area. However, this method requires a large data storage area in order to store the original data and the copy destination data. Therefore, a technique called differential snapshot that manages only data update differences and virtually replicates data is used.

  When using the differential snapshot technology, two data storage areas are set for the storage apparatus. The first data storage area is called a primary volume, and the second data storage area is called a differential volume. The replication source data is stored in the primary volume. Then, when updating the data of the primary volume, the data before the update is saved in the differential volume. When referring to copy destination data, it is first checked whether the corresponding data has been saved in the differential volume. When the data is saved, the difference volume data is referred to. When the data is not saved, the primary volume data is referenced. Further, when the data of the differential volume is updated, the updated data is stored in the differential volume.

Since only the updated data is stored in the differential snap volume, its size is usually small. Therefore, by using the differential snapshot technology, data can be virtually replicated without using a large-size data storage area.
The differential snapshot technique is described in Patent Document 1, for example. Further, Non-Patent Document 1 is known as a technology related to data saving.
JP 2004-272854 A Kazuo Goda, Takayuki Tamura, Masato Oguchi and Masaru Kitsuregawa, `` Run-time Load Balancing System on SAN-connected PC Cluster for Dynamic Injection of CPU and Disk Resource --- A Case Study of Data Mining Application --- '', Proceedings of 13th International Conference on Database and Expert Systems Applications (DEXA2002), September 2002, pp. 182-192

  In a computer system, an application program running on a server device often instructs the storage device to perform data operations. An example of this application program is DBMS (Database Management System). When an application program such as a DBMS updates data, the corresponding data is once read into the memory area of the server device, the data in the memory area is changed, and then the changed data is written out to the storage device.

  Therefore, when updating the data of the primary volume, the primary volume is accessed three times as follows. First, the application program reads data to be updated from the primary volume into the memory area. Next, the application program writes data to the storage device. At the time of writing, the primary volume data is read in order to save the data before update. Also, the updated data is written to the primary volume. Thus, there is a problem that the number of accesses to the primary volume increases and the load on the storage device increases.

  The present invention has been made in view of the above, and an object of the present invention is to reduce the load on a storage apparatus when using a differential snapshot.

One aspect of the present invention is a method for managing data in a storage apparatus including a first volume and a second volume, wherein the server apparatus receives a data operation instruction from a terminal computer, and the server apparatus Determining the operation target data based on the operation instruction for the data, the server device determining whether the operation target data is for update or for reference, and the server device for the server A step of determining whether or not the operation target data is read into a buffer set in the apparatus; the operation target data is for reference; and the operation target data is read into the buffer If not, the server device reads the operation target data read from the first volume into the buffer; When the operation target data is for update and the operation target data is not read into the buffer, the server device stores the operation target data from the first volume to the storage device. Read and issue a first command for instructing to save the operation target data to the second volume, receive the operation target data from the storage device, store the data in the buffer, Executing the update in the buffer, and when the operation target data is for update and the operation target data is read into the buffer, the server device stores the operation target in the storage device. Issue a second command instructing to evacuate the data from the first volume to the second volume, and And executing the update in the buffer, the.

  According to the present invention, when data is read into the memory area, the data can be saved to the differential volume, and the number of accesses to the first volume can be reduced to two. Since the data evacuation has been completed, it is not necessary to read the data of the first volume when updating the data. Thereby, the load of the storage apparatus is reduced.

Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a block diagram showing a configuration of a computer system according to the first embodiment to which the present invention is applied. FIG. 1 shows a computer system in which a DBMS (DataBase Management System) operates on the server 1 and the storage 2 stores the database 101.

  The DBMS 100 of the server 1 receives a data operation request from the operation terminal (terminal computer) 5 or the like via the network 3 and executes data operation on the database 101 according to the request.

  Here, the server 1 is connected to a CPU 11 that executes arithmetic processing, a memory 12 that temporarily stores programs and data, a data transfer controller 13 that controls access to the CPU 11 and I / O, and a network 3. A communication unit (network interface) 14 and a storage interface 15 are provided.

  A program for the DBMS 100 is stored in the memory 12, and the CPU 11 executes the program for the DBMS 100. The data transfer controller 13 transfers data between the CPU 11, the communication unit 14, and the storage interface 15. The storage interface 15 is connected to the storage 2 via a SAN (Storage Area Network) 4. The DBMS 100 includes an SQL execution unit 110, a buffer management unit 120, a disk management unit 130, a data area management table 121, and a DB buffer 140.

  The storage 2 includes a drive 30, a drive 31, and a storage controller 20 that controls these drives. Here, the drive 30 and the drive 31 are devices for storing data. Hereinafter, the drive No. 30 in FIGS. 1 and 2 is referred to as drive 1, and the drive No. 31 as drive 2.

  The storage controller 20 includes a CPU 21 that executes arithmetic processing, a memory 22 that stores programs and data, a data transfer controller 23, an interface 24 connected to the SAN 4, an interface 25 connected to the drive 1, and the drive 2. An interface 26 connected to the network 3 is provided.

  The memory 22 stores programs of the storage control unit 210 and the save instruction determination unit 220, and the CPU 21 executes access to the drive 1 and the drive 2 by executing these programs. The data transfer controller 23 transfers data between the CPU 21 and the interfaces 24 and 25.

  The storage control unit 210 receives a command related to data operation from the server 1 and operates data stored in the drive 1 and drive 2 based on the command. Here, the data operation command is data reference (Read) or data update (Write). However, as will be described later, the storage 2 of the present invention can accept a data reference command including a data saving instruction and a data saving command.

  The storage control unit 210 transmits the received command to the save instruction determination unit 220 in order to determine whether the command received from the server 1 includes a data save instruction. The save instruction determination unit 220 determines whether or not the command includes a data save instruction, and transmits the result to the storage control unit 210. As will be described later with reference to the flowcharts of FIGS. 10 and 12, the storage control unit 210 executes data saving when the command includes a data saving instruction.

  As shown in FIG. 2, the storage control unit 210 manages a volume management table 211. As illustrated in FIG. 5, the volume management table 211 describes the correspondence between volumes and logical data storage areas. This volume may be called a logical unit. In the example of FIG. 5, since “drive 1” is specified in the drive name 2113 of the entry of the volume name 2111 “VOL1”, it indicates that the drive 1 corresponds to VOL1. Similarly, it is shown that the drive 2 corresponds to VOL2. Although it is possible to associate a plurality of drives with one volume, FIG. 5 shows an example in which one drive is associated with one volume.

  In the volume management table 211, a type 2112 and a differential volume name 2114 are described. “Primary” or “Differential” is specified for the type 2112, “Primary” indicates that it is a differential snapshot save source volume, and “Differential” indicates that it is a differential snapshot save destination volume. In the example of FIG. 5, “Primary” is specified for VOL1 and “Differential” is described for VOL2, so VOL1 is the save source volume and VOL2 is specified as the save destination volume. Furthermore, VOL2 is described in the differential volume name 2114 of VOL1, indicating that the save destination volume of the differential snapshot of VOL1 is VOL2. In addition, VOL1 is described in the differential volume name 2114 of VOL2, indicating that the save source volume of the differential snapshot of VOL2 is VOL1.

  The data storage area of the volume is divided into areas of a fixed size such as 512 bytes, and each is assigned an address and managed. This address may be referred to as a logical block address (LBA). FIG. 6 illustrates the database 101 stored in VOL1. The address 1011 is a number assigned to each area of a certain size of VOL1, and in the example of FIG. 6, data “AAA” is stored at address “0” of VOL1. As described above, the drive managed by the storage can be accessed by the volume name and address. For example, when the storage control unit 210 receives a data operation command instructing the read of the data of the address “0” of the VOL 1 from the server 1, “AAA” is transmitted to the server 1.

  Further, the storage control unit 210 manages a difference management table 221 and a save management table 222 as shown in FIG. In the difference management table 221 illustrated in FIG. 7, the address 2211 of the save source volume and the address 2212 of the save destination volume of the differential snapshot are described. The first entry in FIG. 7 indicates that the data at the address “100” of the VOL1 that is the save source volume is saved at the address “0” of the VOL2 that is the save destination volume. In the second and third entries, the data of the VOL1 address “200” is saved to the VOL2 address “1”, and the data of the VOL1 address “300” is saved to the VOL2 address “2”. It is shown that.

  The save management table 222 is a table that records whether the save source volume data has been saved. In the save flag 2222, “0” or “1” is described, “0” indicates that the data has not been saved, and “1” indicates that the data has been saved. In the example of FIG. 8, the save flag 2222 of the address “0” of the VOL 1 that is the save source volume is “0”, indicating that it has not been saved. Similarly, the data of the address “1” and the address “2” of the VOL 1 is not saved. The storage control unit 210 changes the contents of the difference management table 221 and the save management table 222 when the save source volume data is saved to the save destination volume.

  Next, an operation example of the DBMS 100 will be described.

  As shown in FIG. 1, the program of the DBMS 100 is stored in the memory 12 of the server 1 and is executed by the CPU 11. The DBMS 100 receives a data operation request from the operation terminal 5 and creates a query plan corresponding to this request. Here, the data operation request is an instruction to update or refer to data described in SQL (Structured Query Language) or the like. A query plan is a method for executing an instructed data operation, and specifies a volume name and address of data to be read, a method for changing the read data, and the like.

  The SQL execution unit 110 creates a query plan based on the data operation request received from the operation terminal 5. The data operation request is, for example, a reference to data having a table name “Warehouse” and an identification column name “W_id” “1”. In order to generate a query plan from this data operation request, the SQL execution unit 110 uses the data area management table 121. As shown in FIG. 3, the data management table 121 describes a table name 1210, an identification column name 1211, an identification column 1212, a volume name 1213, and an address 1214. In the example of FIG. 3, it can be seen that the data with the table name “Warehouse” and the identification column name “W_id” “1” is stored at the address “0” of the volume name “VOL1”. Accordingly, it is possible to generate a query plan that acquires data of the address “0” of the volume name “VOL1” of the external storage device and transmits the data to the operation terminal 5. In the above, an example in which one piece of data is referred to as a data operation request has been described. However, it is possible to create a query plan for a data operation request including multiple data references and data updates by repeating the above operation. It is.

  The SQL execution unit 110 refers to or updates data according to the generated query plan. In the case of data reference, the SQL execution unit 110 notifies the buffer management unit 120 of the volume name and address of the data to be referenced. The buffer management unit 120 manages the DB buffer 140, and when the requested data is in the DB buffer 140, passes this data to the SQL execution unit 110. The SQL execution unit 110 transmits the data received from the buffer management unit 120 to the operation terminal 5. The contents stored in the DB buffer 140 are illustrated in FIG. In the DB buffer 140, a volume name 1401, an address 1402, and data 1403 are described, and it can be determined whether or not the data requested from the SQL execution unit 110 has already been read into the DB buffer 140. When the data requested from the SQL execution unit 110 is not read into the DB buffer 140, the buffer management unit 120 issues a data reference command to the storage 2 and stores the read data in the DB buffer 140. Then, the buffer management unit 120 passes the data to the SQL execution unit 110.

  When the SQL execution unit 110 updates data, the buffer management unit 120 is notified of the volume name and address of the data to be updated. If the requested data is in the DB buffer 140, the buffer management unit 120 passes this data to the SQL execution unit 110. When the requested data is not in the DB buffer 140, the buffer management unit 120 issues a data read command to the storage 2, stores the read data in the DB buffer 140, and passes the data to the SQL execution unit 110. . The SQL execution unit 110 updates the received data and passes the data to the buffer management unit 120. The buffer management unit 120 stores the received data in the DB buffer 140 and issues a data update command to the storage 2.

  In the operation example of the DBMS 100 described above, the SQL execution unit 110 can determine whether the data is for reference or for update. For example, when a data update request is received from the operation terminal 5, the data is update data. Further, when a query plan including a plurality of data references and data updates is generated, if the referenced data is updated later, it can be determined that the data is updated at the time of data reference. If it is determined that the data is update data, the SQL execution unit 110 notifies the buffer management unit 120 that the data is a volume name, an address, and update data. The buffer management unit 120 notified of the update data transmits a data reference command to the save instruction adding unit 130. The save instruction adding unit 130 adds a tag for instructing data save to the data reference command, and transmits the tag to the buffer management unit 120. The buffer management unit 120 transmits to the storage 2 a data reference command to which an instruction to save data is added. The storage 2 that has received the data read command with the data save tag saves the data of the save source volume to the save destination volume, and transmits the data to the server 1.

  Further, the buffer management unit 120 can issue a command for instructing only saving to the storage 2 when update data already exists in the DB buffer 140. As a result, the storage 2 can execute in advance the operation of saving the data of the save source volume to the save destination volume.

  In addition to the above-described embodiment, there is an embodiment in which the storage 2 manages the save volume setting table illustrated in FIG. In this embodiment, when the volume data set in the save volume setting table is accessed, the storage 2 saves the data to the save destination volume.

<Details of processing>
Next, details of processing performed by the DBMS 100 of the server 1 and the storage control unit 210 of the storage 2 will be described. FIG. 9 is a flowchart illustrating an example of processing performed by the DBMS 100 and is executed every time a query is received from the operation terminal 5.

  When the DBMS 100 receives a query (data operation instruction) from the operation terminal 5 (S1), the DBMS 100 analyzes the contents of the query and creates a query plan (S2). In S3, data to be read from the storage 2 to the server 1 is determined based on the created query plan.

  In S4, it is determined whether or not data to be read into the server 1 has been read into the DB buffer 140. If the data to be read has already been read into the DB buffer 140, the process proceeds to S9. If the corresponding data is not read in the DB buffer 140, the process proceeds to S5.

  In S5, it is determined whether or not the data read from the storage 2 is update data. That is, the query plan created in S2 is analyzed, and it is determined whether or not to update after reading the corresponding data into the server 1. If it is update data, the process proceeds to S6 to generate a reference command with a data save tag. If it is not update data, a normal reference command is set and the process proceeds to S7.

  In S7, the reference command or reference command with the data save tag set in S5 and S6 is transmitted to the storage 2.

  In S8, a reference command with a data save tag or data from the storage 2 for the reference command is received. Then, the DBMS 100 stores the received data in the DB buffer 140.

  Next, in S9, it is determined whether or not to update the data read (or already read) into the DB buffer 140. When updating, it progresses to S10 and the update according to a query plan is performed with respect to the applicable data on DB buffer 140. FIG. On the other hand, if the read data is not updated, the process proceeds to S11. In S11, the result of the query is notified to the operation terminal 5 that has transmitted the query.

  When the data read into the DB buffer 140 of the server 1 is updated by executing the above processing every time a query is received from the operation terminal 5, a reference command with a data save tag is issued to match the data reference. The corresponding data can be saved.

  The data operation instruction transmitted from the operation terminal 5 to the server 1 in S1 is, for example, “UPDATE Warehouse SET name =“ Tokyo ”WHERE W_id = 1”. This is an instruction to change the value of the name column 1212 in which the identification column name 1211 called W_id of the table name 1210 called Warehouse is “1” to Tokyo as shown in FIG.

In the above, an example of the UPDATE statement is shown, but a data operation instruction such as “SELECT name FROM Warehouse Where W_id = 1” can also be accepted. This is an instruction to read data in the name column whose W_id column in the Warehouse table is “1”.
Next, processing performed by the storage controller 20 of the storage 2 will be described. 10, 11, and 12 are flowcharts illustrating examples of processing performed by the storage controller 20. FIG. 10 shows processing when the storage controller 20 receives a reference command or a reference command with a data save tag from the SAN 4. FIG. 11 shows processing when the storage controller 20 receives a write command from the SAN 4, and FIG. 12 shows processing when the storage controller 20 receives a data save command from the SAN 4.

  First, in the process when the reference command or the reference command with data save tag in FIG. 10 is received, the storage controller 20 receives the reference command (reference command with tag) in S21. Next, in S22, the storage controller 20 reads data at the address specified by the reference command from the primary volume.

  Next, in S23, the tag of the received reference command is referred to and it is determined whether or not it is a tagged reference command. For example, as shown in FIG. 13, when a command received by the storage controller 20 is represented by n bits, 0 bit is set as a tag area and 1 to n bits are set as a command area. For example, when the value of the tag area is “1”, an instruction to save data is indicated, and when the value is “0”, no instruction to save data is indicated.

  In S23, the presence / absence of a data saving instruction is determined based on the value of the tag area added to the reference command. If it is determined that there is an instruction, the process proceeds to S24 to execute data saving. On the other hand, if it is determined that there is no instruction, the process proceeds to S27 to transmit this data to the requesting server 1.

  In S24 in the case of a reference command with a data save tag, the storage controller 20 refers to the save management table 222 for the address of the primary volume read in S22, and whether save has been completed based on the value of the save flag 2222. Determine whether. If the result of this determination is that the corresponding data on the primary volume has not been saved, the process proceeds to S26 to create a differential snapshot. On the other hand, if the corresponding data has been saved, the process proceeds to S27.

  In S26, the data read from the primary volume is copied to the differential volume to save the data. Then, when the data saving is completed, the storage controller 20 writes the copied primary volume address 2211 and the saving destination address 2212 of the saving destination differential volume in the difference management table 221. Further, since the data saving is completed, the saving flag 2222 of the saving management table 222 is set to “1” for the corresponding address 2221 on the primary volume, and the data saving completion is recorded.

  In S <b> 27, the storage controller 20 transmits the data whose data has been saved to the requesting server 1 and ends the processing. In the case of a reference command without a data save instruction, the primary volume data read in S22 is transmitted to the server 1 as it is.

  Thus, in the case of a reference command with a data save tag, the data read by the storage controller 20 is copied from the primary volume to the differential volume, and then the data requested by the server 1 is transmitted. When the data saving is completed, the difference management table 221 and the saving management table 222 are updated.

  Next, processing when the storage controller 20 receives an update command will be described with reference to FIG.

  First, in S31, the storage controller 20 receives an update command. Next, in S32, the storage controller 20 determines whether or not the data at the address to be written has been saved. If the result of this determination is that the save has not been completed, the process proceeds to S33, and the data at the address of the write target primary volume is saved in the differential volume. When the data save is completed, the copied primary volume address 2211 and the save destination address save destination address 2212 are written in the difference management table 221.

  On the other hand, if it is determined in S32 that the data of the primary volume to be written has already been saved, the process proceeds to S34.

  In S34, the storage controller 20 updates the data of the write target primary volume. In S35, the storage controller 20 sends a notification that the data has been written to the server 1 and ends the process.

  Next, processing performed when the storage controller receives a data save command will be described with reference to FIG.

  First, in S41, the storage controller 20 receives a data save command. Next, in S42, the storage controller 20 determines whether or not the saving of the data at the address on the data saving target primary volume has been completed in the same manner as in S24 and S25. If the result of this determination is that saving has not been completed, the process proceeds to S43 to read the data of the primary volume to be written. In S44, the read data is copied to the differential volume and the saving is completed. When the saving of the primary volume data is completed, the storage controller 20 writes the copied primary volume address 2211 and the saving destination address saving destination address 2212 in the difference management table 221. Further, the save flag 2222 of the corresponding address 2221 in the save management table 222 is reset to “1”, and the fact that the save has been completed is recorded.

  On the other hand, if the data of the primary volume to be written has already been saved, the process proceeds to S45.

  In S <b> 45, the storage controller 20 transmits a notification that data saving has been completed to the server 1 and ends the process.

  As described above, in the case of the data save command, the storage controller 20 saves the data of the primary volume to be saved to the differential volume, and updates the difference management table 221 and the save management table 222.

  The data save command may be issued by the DBMS 100 to the storage 2 when the data to be referred to by the DBMS 100 has already been read into the DB buffer 140 and has been updated on the DB buffer 140. good. In this case, when there is a reference request for data that has been updated in the DB buffer 140 and has not been written to the storage 2, the storage controller 20 of the storage 2 accepts the data save command and before the writing occurs In addition, a differential snapshot can be created in advance. As a result, when the data updated on the DB buffer 140 is updated in the storage 2, a differential snapshot has already been created. Therefore, only the writing needs to be performed, and the storage 2 can be processed quickly. it can.

  As described above, in the present invention, when the server 1 references data to the storage 2, it is determined whether or not the data to be referenced is accompanied by an update. Issue a reference command with a tag that indicates saving. The storage controller 20 of the storage 2 that has received the reference command with the data save tag transmits this data to the server 1 after the save of the data requested for reference is completed. As a result, data can be saved and data can be read only by transmitting the tagged reference command from the server 1 to the storage 2 once.

  In the above-described embodiment, an example in which the DBMS 100 is operated on the server 1 has been described. However, when requesting data reference to the storage 2, an application that can determine or estimate whether or not to update the referenced data later The present invention can be applied to any service.

  Moreover, although the example in which the server 1 and the storage 2 are connected by the SAN 4 is shown, the present invention can be applied even when the storage 2 is connected to the server 1 through an interface such as SCSI or SATA.

Second Embodiment
FIG. 14 shows the data to be updated even when a part of the processing of the DBMS 100 running on the server 1 shown in FIG. 9 of the first embodiment is changed and the data requesting reference exists in the DB buffer 140. If there is, a tagged reference command is executed, and data is saved in advance. Other configurations are the same as those in the first embodiment.

  In FIG. 14, the processing of the DBMS 100 is the same as the processing of S1 to S3 and S5 to S11 in FIG. 9 of the first embodiment, and the case where the determination of S4 is YES is FIG. 9 of the first embodiment. And different.

  In S4, if there is no data to be referenced in the DB buffer 140, the process proceeds to S5 as in FIG. 9 of the first embodiment. On the other hand, if the data to be referenced already exists in the DB buffer 140, the process proceeds to S51 to determine whether the data to be updated. The query plan created in S2 is analyzed, and it is determined whether or not to update after reading the corresponding data into the server 1. In the case of updating, the process proceeds to S52 to set a tagged reference command in order to create a differential snapshot. In S52, a tagged reference command is set for the data in the DB buffer 140, and the process proceeds to S7 to instruct the storage 2 to refer to the data and create a differential snapshot.

  On the other hand, if it is not necessary to update in the determination of S51, the process proceeds to S11 to notify that it has already been read.

  As described above, even when data to be read by the DBMS 100 exists on the DB buffer 140, if the data is updated to the server 1 or is expected to be updated after the data is read to the server 1, the tagged reference command is used. Issue a differential snapshot of the data. As a result, when the DBMS 100 updates the data that is expected to be updated and requests writing to the storage 2, the data is already written because the creation of the differential snapshot has already been completed in advance. It will be. As a result, in the storage 2, it is not necessary to create a differential snapshot for the data requested to be written, and the writing process can be speeded up.

<Third Embodiment>
In addition to the above embodiment, there is an embodiment in which the storage controller 20 manages the save volume setting table illustrated in FIG. The storage controller 20 stores the differential save attribute table 223 in the memory 22. The save volume setting table 223 stores the volume name for saving the referenced data to the differential volume. The difference saving attribute table 223 is set from the management terminal 6 or the like.

  When the storage control unit 210 receives a reference command from the server 1, the storage control unit 210 determines whether the volume specified in the command is set in the save volume setting table 223. If the volume specified in the command is set in the save volume setting table 223, the data of the primary volume is saved in the differential volume, and then the data is transmitted to the server 1.

  Even in such an embodiment, it is possible to reduce the load on the primary volume by saving the data of the primary volume to the differential volume at the time of data reference.

  In this specification, the DBMS is used as an application program for accessing data, but the application program is not limited to the DBMS.

    As described above, according to the present invention, it is possible to reduce the load on the storage apparatus by saving data when referring to data.

The block diagram of the computer system which shows the 1st Embodiment of this invention. FIG. 2 is a functional block diagram of a server DBMS and a storage control unit of an external storage device. Explanatory drawing which shows an example of a data area management table. Explanatory drawing which shows an example of DB buffer. Explanatory drawing which shows an example of a volume management table. Explanatory drawing which shows the content of a database. Explanatory drawing which shows an example of a difference management table. Explanatory drawing which shows an example of an evacuation management table. The flowchart which shows an example of the process performed by DBMS of a server. The flowchart when an example of the process performed by the storage control part of an external storage device is shown, and a reference command or a tagged reference command is received. The flowchart when an example of the process performed by the storage control part of an external storage device is shown, and the write command is received. The flowchart when an example of the process performed in the storage control part of an external storage device is shown, and the data saving command is received. Explanatory drawing of a command and a tag. The flowchart which shows 2nd Embodiment and shows an example of the process performed by DBMS of a server. Explanatory drawing which shows 3rd Embodiment and shows an example of the difference save attribute table managed by the storage control part of an external storage device.

Explanation of symbols

1 server 2 storage 20 storage controller 30 drive 1
31 Drive 2
100 DBMS
101 Database 102 Difference Data 110 SQL Execution Unit 120 Buffer Management Unit 130 Save Instruction Addition Unit 140 DB Buffer 210 Storage Control Unit 220 Save Instruction Determination Unit 211 Volume Management Table 221 Difference Management Table 222 Save Management Table

Claims (4)

A method for managing data in a storage device comprising a first volume and a second volume,
  A step in which the server device receives a data operation instruction from the terminal computer;
  The server device determining operation target data based on an operation instruction of the data;
  The server device determining whether the operation target data is for update or for reference; and
  The server device determining whether or not the operation target data is read in a buffer set in the server device;
  When the operation target data is for reference and the operation target data is not read into the buffer, the server device reads the operation target data read from the first volume. Reading into the buffer;
  When the operation target data is for update and the operation target data is not read into the buffer, the server device stores the operation target data from the first volume to the storage device. Read and issue a first command for instructing to save the operation target data to the second volume, receive the operation target data from the storage device, store the data in the buffer, Performing an update in the buffer;
  When the operation target data is for update and the operation target data is read into the buffer, the server device stores the operation target data in the storage device from the first volume. Issuing a second command instructing to save to the second volume, and executing an update in the buffer;
  Data management method for storage device including   The data operation instruction received by the server device from the terminal computer includes an SQL statement,
  The step in which the server device determines whether the operation target data is for update or for reference is determined according to the type of SQL sentence included in the operation instruction.
  The data management method for a storage apparatus according to claim 1.   A computer system including a storage device having a first volume and a second volume, and a server device,
  The server device receives data operation instructions from a terminal computer,
  The server device determines operation target data based on an operation instruction of the data,
  The server device determines whether the operation target data is for update or reference;
  The server device determines whether or not the operation target data is read in a buffer set in the server device,
  When the operation target data is for reference and the operation target data is not read into the buffer, the server device reads the operation target data read from the first volume. Read into the buffer,
  When the operation target data is for update and the operation target data is not read into the buffer, the server device stores the operation target data from the first volume in the storage device. Read and issue a first command for instructing to save the operation target data to the second volume, receive the operation target data from the storage device, store the data in the buffer, Perform updates in the buffer,
  When the operation target data is for updating and the operation target data is read into the buffer, the server device stores the operation target data in the storage device from the first volume. Issue a second command instructing to save to the second volume, execute an update in the buffer,
  The storage apparatus that has received one of the first command and the second command determines whether or not the operation target data has been saved according to the one command. According to one command, the operation target data is saved from the first volume to the second volume.
  Computer system.   The data operation instruction received by the server device from the terminal computer includes an SQL statement,
  The server device determines whether the operation target data is for update or for reference according to the type of SQL statement included in the operation instruction.
  The computer system according to claim 3.

Images